Machine Studying (ML), AI, Pc Imaginative and prescient (CV)… you’d be arduous pressed to seek out somebody who hasn’t a minimum of heard of those subjects. They dominate standard tradition and are on the forefront of a brand new group of technological potentialities. A wonderful instance is the Meraki MV digicam – a digicam that stretches the definition of ‘digicam’. Powered by ML and CV, the MV is a visible sensor, providing vital insights about individuals and autos in your bodily areas. These insights give attention to the counting & movement of objects in body at a specific time, and they’re relevant to a wide range of use instances.
Whereas many challenges might be addressed with the MV’s native capabilities, what if our use case is extra advanced? We is perhaps occupied with detecting topics aside from individuals or autos. Maybe we have to classify a car by make & mannequin, or classify an individual by age. How will we deal with these use instances?
One technique to resolve these challenges is to pair the MV with a customized ML mannequin. We will leverage the MV’s snapshots, carry out picture evaluation utilizing our ML mannequin, and act on the outcomes. Another choice is the brand new “Customized Pc Imaginative and prescient” function. This function permits us to insert our personal ML mannequin immediately onto the digicam, subscribe to a MQTT matter, and extract classification knowledge. What would these concepts seem like in apply? Nicely let’s have a look at actual examples of the place our group carried out these concepts to assist Cisco clients resolve distinctive challenges. Every case is exclusive, however all of them share frequent parts: an MV Digicam, a customized ML Mannequin, and a little bit little bit of customized code.
Individuals Detection 2.0
On this instance, we labored with a retail buyer occupied with creating a personalised purchasing expertise utilizing digital signage and Meraki MV’s. Our group used a mix of MV Individuals Detection, Python, and AWS’s Rekognition Engine to boost the shopper purchasing expertise. The answer detects buyers, identifies their age & intercourse, and makes use of digital signage to show knowledge pushed, related product strains. The answer not solely improved the shopper’s expertise however result in elevated gross sales and additional analytics round buyer foot visitors and demographics. Take a look at the code right here.
Car Detection 2.0
On this instance, we labored with a buyer occupied with bettering their curbside pickup expertise by lowering buyer wait occasions. Our group constructed an answer that detected a automobile, matched the license plate to an order, and despatched a Webex notification informing staff a buyer with the corresponding order was ready. Powered by Flask, MV Car Detection, and the Google Imaginative and prescient Engine for Plate Detection, the answer considerably decreased curbside pickup time and elevated worker effectivity. Take a look at the code right here.
Past Individuals and Vehicles
In our ultimate instance, a buyer wanted to simply implement masks utilization in a big public venue throughout Covid-19. Our resolution leveraged the MV’s RTSP Video Feed, Python, and a customized TensorFlow Masks Mannequin to investigate MV footage for masks utilization. If an individual in body was not sporting a masks, the code sends a Webex alert and snapshot to safety personnel. This resolution helped the venue meet compliance and security restrictions through the pandemic. Take a look at the code right here.
Cameras & ML Fashions vs. The World
The Meraki MV Cameras supply a robust platform for automation, however generally a use case goes past native capabilities. Happily, with a little bit of customized code & a customized mannequin, it’s simpler than ever to increase an MV and deal with a buyer’s most tough challenges.
For those who’re occupied with studying extra in regards to the examples, try the hyperlinks beneath. Every repository incorporates the pattern code & directions for easy methods to use it in your personal community:
About our GVE group
The International Digital Engineering (GVE) DevNet group works with Cisco clients to assist deliver their automation concepts to life. Along with Cisco Account Groups, we discover alternatives the place clients want a little bit assist getting began with automation or integration initiatives. We develop easy examples to showcase what is feasible with a little bit little bit of customized code. Many of those instance initiatives are revealed on the GVE DevNet GitHub web page and shared with the neighborhood.
Associated sources
We’d love to listen to what you assume.
Ask a query or depart a remark beneath.
And keep linked with Cisco DevNet on social!
LinkedIn | Twitter @CiscoDevNet | Fb | YouTube Channel
Share:
